has joined #pact-ruby

/github subscribe pact-foundation/pact-ruby

/github unsubscribe pact-foundation/pact-ruby deployments public commits

/github subscribe pact-foundation/pact-mock_service

/github unsubscribe pact-foundation/pact-mock_service commits

/github unsubscribe pact-foundation/pact-mock_service public

/github unsubscribe pact-foundation/pact-mock_service deployments

/github subscribe pact-foundation/pact-support issues pulls releases statuses

/github unsubscribe pact-foundation/pact-support public

/github unsubscribe pact-foundation/pact-support commits

/github unsubscribe pact-foundation/pact-support deployments

/github unsubscribe pact-foundation/pact-support statuses

/github unsubscribe pact-foundation/pact-mock_service statuses

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Hi, I am trying to figure out how to do file_upload in ruby. Both provider and consumer are written in ruby. I did see pact-jvm examples with scala. But really need examples of how i can set it up in ruby

Hi @nabrosimova. What content type are you using?

hi, i have a file upload call that in place of body has payload something like this `multipart: true, file: file, name: name`

the whole problem is.. it first doesnt see the payload. second of all i need to pass in file for the call and i dont know how to do that in ruby. i saw this https://github.com/DiUS/pact-jvm/commit/3f9a03ef66d97f411d2b959527a693f2c2cd188a but couldnt find an example of how to use it in ruby

when i check how the call actually looks like it looks something like "headers": { "Content-Type": "multipart/form-data; boundary=----WebKitFormBoundary1bvPHrYmtx0UNv2e" }, "body": "------WebKitFormBoundary1bvPHrYmtx0UNv2e\r\nContent-Disposition: form-data; name=\"file\"; filename=\"476.csv\"\r\nContent-Type: text/csv\r\n\r\ndata\r\n------WebKitFormBoundary1bvPHrYmtx0UNv2e--\r\n" }

but RubyFormBoundaryPs4ifEBhAkIzfpNe is autogenerated and is always different... so i cannot just make it up

@bethskurrie is there any way i can test upload file without having to worry about rubyForm boundary changing

@uglyog I'm sure I've seen an example for pact-jvm with this. How did you get around it?

It wouldn't be too hard to write a diff for multipart form, but I want to keep the functionality as similar as possible.

Oh, right, just read the groovy link. I'll see if I can work it out when I'm at my computer.

has joined #pact-ruby

I added support for it in the DSL: https://github.com/DiUS/pact-jvm/blob/master/pact-jvm-consumer-junit/src/test/groovy/au/com/dius/pact/consumer/junit/ExampleFileUploadSpec.groovy

So, parse the body and do the matching ignoring the boundary.

Exact matching for the non boundary fields?

yeah i tried to do this manually

body: Pact.like('------RubyFormBoundary2AQPyCUtPGtlPQKQ\r\nContent-Disposition: form-data; name=\"file\"\r\n\r\n\r\n------RubyFormBoundary2AQPyCUtPGtlPQKQ\r\nContent-Disposition: form-data; name=\"name\"\r\n\r\nFish in a Pond!\r\n------RubyFormBoundary2AQPyCUtPGtlPQKQ--\r\n')

but the problem is it doesnt see it as file, it sees it as string

and i am not sure how to use groovy with ruby

"headers": { "Content-Type": "multipart/form-data; boundary=----RubyFormBoundary2AQPyCUtPGtlPQKQ", "Content-Length": "235", "Accept": "*/*", "Accept-Encoding": "gzip, deflate", "User-Agent": "rest-client/2.0.2 (linux-gnu x86_64) ruby/2.4.4p296", "Authorization": "pact_notorious_secret", "HOST": "localhost:2345" }, "body": "------RubyFormBoundary2AQPyCUtPGtlPQKQ\\r\\nContent-Disposition: form-data; name=\\\"file\\\"\\r\\n\\r\\n\\r\\n------RubyFormBoundary2AQPyCUtPGtlPQKQ\\r\\nContent-Disposition: form-data; name=\\\"name\\\"\\r\\n\\r\\nFish in a Pond!\\r\\n------RubyFormBoundary2AQPyCUtPGtlPQKQ--\\r\\n",

@nabrosimova you won't be able to do it until I write a new feature for pact ruby.

Can you raise an issue in the pact-ruby repository please?

i did raise an issue but also i got it to almost work.. though it gives me `Encoding::UndefinedConversionError - "\x9F" from ASCII-8BIT to UTF-8` error

when i try to make pact with real non-empty file

also, i can try to make pact with empty mov file but it fails on returning me the wrong response, by saying that transcoding process failed

Also, i am just an intern and my time ends soon, so I was wondering if that feature will be added, how long it will take?

It shouldn't take too long, but my time is limited. I'll see if I can get it done this week. Is that fast enough?

0.0 that fast wow!

i will try to help you as much as I can, just let me know what you needf

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Is it possible to have a pact spec that tests multiple requests?

That is a very ambiguous question @lkingsley!

Do you mean, can you set up multiple interactions on the mock service at the same time? Yes.

Do you mean, can you chain together a group of requests to replay together? No.

Do you mean, can you mock multiple requests to different providers at the same time? Yes.

More specifically I want to upload a file but first need to send a post request telling the server about the size and file type so it can be ready for it. It will then return a unique key, a new url for me to send another post request to but Im unsure how to set up the test if I do not know the url or key value until the first test has run. Does this sound at all possible? @bethskurrie

No. You need to test each request separately

Use provider states to set up the correct data so that when you do the file upload, it matches the file.

Okay thanks. Sorry for the confusion.

Np

There is a section on provider states in http://docs.pact.io if you're unfamiliar with them.

Yea the only issue is when setting this up in a provider state it doesn?t give me the necessary information to make the request.

has joined #pact-ruby

hey, i am running into weird error trying to implement verification in jenkins...

it looks like gem cannot find spec_helper?

has joined #pact-ruby

Hello! Quick question! Is there a way to consume `params` from a provider state listed in a pactfile while implementing the `provider_state` on the provider? _Example:_ Pact file: ``` "interactions": [ { "description": "/api/v3/animals/:id", "providerStates": [ { "name": "an animal named Spot", "params": { "name": "Spot" } } ], ``` Provider state: ``` provider_state "an animal named Spot", params do Dog.new(params.name) end ``` Apologies if any of this is unclear, I'm happy to clarify just let me know! Thanks!!

Not yet @tom397 because I haven't yet implemented v3 pact specification in Ruby.

If you're using param, your provider is v3, and ruby only implements v2 so far (just because I haven't had time to do it)

@tom397 I've added provider state params support to the ruby impl. ``` set_up do | params | do Alligator.new(name: params.fetch('name')).save! end ```

You'll need version 1.36.0 of the pact gem

if you're verifying v3 pacts, you can now access the provider state params in the set up and tear down blocks. You'll need version 1.36.0 of the pact gem. ``` set_up do | params | do Alligator.new(name: params.fetch('name')).save! end ```

has joined #pact-ruby

Amazing!! Thanks @bethskurrie!

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

@bethskurrie I found an old discussion https://github.com/pact-foundation/pact-ruby/issues/49, do we have any system in place generate a valid JWT:Token in the `Provider::Verify`? I have a `before_action :ensure_signed_jwt!` in my controller with a JWT valid for 10 minutes.

has joined #pact-ruby

As a general question, how the provider proxy relates to the the actual provider?

Do I have to use a brand new application as Pact Provider proxy?

Not yet, but you can use middleware to modify the request with a live token. Which language are you using @marco.dellolio?

For the moment ruby, but I will have the same issue with Elixir

and potentially a Node application

Do you mean in the ruby? It's for performing verifications against a running provider, rather than a rack app.

I was looking into the https://github.com/pact-foundation/pact-provider-verifier is it a standalone application? And more specifically how it relates to the real provider? I?m not quite sure about the workflow

With native ruby, make some rack middleware and wrap it around your app when you configure it on the provider side.

You'll find an example in the "verifying pacts" documentation on the Pact ruby wiki.

With the other languages, it's the same middleware, but you'll configure it differently. Look at the docs for the pact provider verifier cli.

Pact provider verifier is for use against running apps of any language. It's packaged as a cli, but if you're using ruby, you'd just use the native gem. It's much more elegant.

It's deprecated now, in favour of the pact provider verifier

Yeah I?ve been reading those today. If I understand correctly the flow is the following. I put my pact files in the proxy provider (locally or broker), the proxy provider is a standalone app (a docker container e.g) which hits the real provider end-point and verify the pact file?

There is no proxy.

The pacts go in the broker.

The broker does not proxy, and it does not run any tests. It is a repository for pacts and verification results.

The consumer tests run with a mock service and generate the pacts

The pact is verified against a rack app, if you're using native ruby, or a running application if not.

The pact can be exchanged via a broker, or if you're just playing around, you can just point to the local file on your file system.

The pact provider verifier takes the pact, and behaves like a mock consumer, replaying the request against the provider, and checking the response matches the expected response.

until this point it?s all clear. What I still not fully understand is : Where https://github.com/pact-foundation/pact-provider-proxy/ this should live? In my provider or it is an independent application?

Ignore that. You don't need it.

ok

We don't use that any more.

It's been replaced with the pact-provider-verifier

But if you're verifying a ruby rack app, you just need the pact gem, nothing else.

In a ruby app are you actually suggesting this option? ``` class ProxyApp def initialize real_app @real_app = real_app end def call env @real_app.call(env.merge('HTTP_AUTHORIZATION' => '12345')) end end Pact.service_provider "Some Provider" do app do ProxyApp.new(RealApp) end honours_pact_with "Some Consumer" do #... end end ```

Yup

I?m making my way through it! The problem is the new application doesn?t recognize my models. When I run this in the provider state ``` set_up do raw_email = File.read("spec/fixtures/emails/email.eml") RawEmail.create( tenant_id: 1, message_id: "CAAWRWJD5Dvw64zAzYurfsfDB22UGakRBd=P0ZCmjV7_UZPuxDg@mail.gmail.com", raw_email: raw_email ) end ``` it errors out `NameError:uninitialized constant RawEmail`

Should I manually import my models?

Is this a rails app?

yes

What is the main entry point of your app if you were running it? The thing that requires all the other files? You'll need to require that in the pact helper.

I?m already requiring `'config/application'` which already requires `require "rails/all"` and `config.autoload_paths += [Rails.root.join("app")]` which includes the models

I'm afraid I can't help you without seeing your code then. It sounds like it should work.

It runs OK when you use the real endpoint?

yep

has joined #pact-ruby

has joined #pact-ruby

where should I start to look into to start implementing V3 matchers?

Start with https://github.com/pact-foundation/pact-specification/tree/version-3#allow-arrays-of-matchers-to-be-defined-against-a-matcher-path which describes the changes from V2

Then have a look at the existing implementations in Pact-JVM and in the Rust implementation

Ok, thank you

Hey all, I?m having hard time parsing my provider verifier query params in a rails app

this is how the request looks like in the controller ``` request => #<ActionDispatch::Request:0x00007fbb11d3cb50 @env= {"rack.version"=>[1, 3], "rack.input"=>#<StringIO:0x00007fbb133b93d8>, "rack.errors"=>#<StringIO:0x00007fbb133b9568>, "rack.multithread"=>true, "rack.multiprocess"=>true, "rack.run_once"=>false, "REQUEST_METHOD"=>"GET", "SERVER_NAME"=>"http://example.org", "SERVER_PORT"=>"80", "QUERY_STRING"=>"%7B%22tenant_id%22%3A1%2C%22message_id%22%3A%22CAAWRWJD5Dvw64zAzYurfsfDB22UGakRBd=P0ZCmjV7_UZPuxDg%40mail.gmail.com%22%7D", ```

this is how params get parsed `<ActionController::Parameters {"{\"tenant_id\":1,\"message_id\":\"CAAWRWJD5Dvw64zAzYurfsfDB22UGakRBd"=>"\"}", "format"=>"json",`

So the problem is PACT is building the query string out ``` "body": { "tenant_id": 1, "message_id": "" } ```

parsing the object and not the key values resulting in this ``` "%7B%22tenant_id%22%3A1%2C%22message_id%22%3A%22CAAWRWJD5Dvw64zAzYurfsfDB22UGakRBd=P0ZCmjV7_UZPuxDg%40mail.gmail.com%22%7D" ``` instead of ``` tenant_id=1&message_id=22CAAWRWJD5Dvw64zAzYurfsfDB22UGakRBd=http://P0ZCmjV7_UZPuxDg%40mail.gmail.com ```

has joined #pact-ruby

has joined #pact-ruby

Hey, I just stumbled upon the following scenario: I started implementing pact on an existing project that currently uses VCR. I was getting errors from VCR about unrecorded cassettes for requests that pact was trying to make to the mock service. Obviously I don?t want to record those. The solution was to add the following configuration to VCR to ignore requests made to the mock service: ``` VCR.configure do |c| config.ignore_request do |request| URI(request.uri).port == 1234 # Ignore requests to Pact mock service end end ``` I would like to add this little piece of information to the docs somewhere, to save some research time for future users. Where would be the best place for this? Is there a ?gotchas? section I can put this under?

has joined #pact-ruby

Has anyone here tried using pact with protobufs?

@bethskurrie?

Hi :wave: I?m using pact-ruby to write a contract test for our consumer and provider service. Our services communicate via protos and the protos are encoded when sent over the wire. When the protos are encoded, they contain certain characters that are not JSON serializable (pact throws; an error I received looks like this - "\xC8" from ASCII-8BIT to UTF-8). I was able to overcome this by Base64 encoding the request and response - in the pact gem - before it is JSON serialized in the consumer - and then Base64 decoding it in the provider before it is used for verification. I?m wondering if there?s a better solution for what I?m doing. Can someone help?

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Hey guys, quick question, should I push the automatically created pact json file to the git repository?

has joined #pact-ruby

While debugging my unit test on the consumer, is there a way to list all the interactions currently registered on the mock service?

has joined #pact-ruby

Hey guys, are you going to resolve problem with encoding https://stackoverflow.com/questions/51604866/pact-verifier-is-failing-with-encodingundefinedconversionerror-xe2-from-as https://github.com/pact-foundation/pact-python/issues/97 and https://github.com/pact-foundation/pact-ruby-standalone/issues/27 I have the same problem with encoding on pact js when im trying to post request to my mock. Returned ruby error stack with ` Error ocurred in mock service: Encoding::UndefinedConversionError - "\x89" from ASCII-8BIT to UTF-8`

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Hey, I am currently working on a project that requires a lot of key values to be compared in response where response is a hash, i want to use something like `Pact.like(responseHash)` instead of `Pact.like(key1: "value1", Key2: "value2" ,...)`, but the finally created json `"matchingRules": { "$.body": { "match": "type" } }`

instead of matching each key value pair

GOLang version of has something like dsl.match which can take in a struct

do we have a similar thing in pact-ruby as well?

can someone share the tentative date of v3 release?

@000sps97 both snippets of code you mentioned will work the same. The resulting matching rule will match each key value pair in the `responseHash` you provide as you want it to

@simon.nizov does ` matchingRules": { "$.body": { "match": "type" } } ` match to every key value in the response ? shouldn't it be something like ` matchingRules": { "$.body.key1":{"match":"type"} "$.body.key2":{"match":"type"} }

@000sps97 yup. You can test it easily by verifying the pact against your provider

Is there a way i can use TEST_DB on provider side than using Development DB ?

If you run rspec on test dB you can just do ?require rails_spec.rb? which fixes this

This can be ignored

has joined #pact-ruby

Hello! I've looked online and can't find anybody mentioning this use-case, so was wondering if anybody here had any suggestions. I'm building a collection of rails apps, which each have an internal API provided by a Rails Engine (gem) that I'm building. That API is both a consumer and provider, and is tested using Pact. I've got the consumer tests working fine, and verified on the provider (another rails app). However, I'm struggling to get the provider verification working in the Rails Engine.

because it's not a full Rails app, there is no `http://config.ru` in the root of the project

I get this when trying to verify the pacts: ```Failure/Error: raise "Could not find http://config.ru file at #{Pact.configuration.config_ru_path} Please configure the service provider app or create a http://config.ru file in the root directory of the project. See https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts for more information." ```

Is there a rack app?

Well, there's the dummy app in the `spec` folder used for testing

I tried setting up the verify task in the Rakefile for the dummy app, but that wasn't happy either

I don't know what a rails engine is, but if there is a rack interface, you can just set the app { RailsEngine.new } in the top level of the provider config.

Sorry, on phone.

Rails Engine is basically a mini rails app that you can mount inside another rails app

definitely rack based

Right.

i'll give that a try

So, set the app. You should find docs in the Verifying Pacts page.

Of the wiki.

Ah ... fixed it

I had to add this in my pact_helper:

```Pact.configuration.config_ru_path = "spec/sample_app/config.ru" ```

:+1:

thank you!

NP.

BTW, @jon.pascoe I thought I'd fixed the interpolation in that error message you got. What version of the pact gem are you using?

`pact (1.40.0)`

`pact_broker-client (1.18.0)`

I think they?re the most recent versions @bethskurrie

On the topic of `webmock` ? I have an issue with it interfering with uploading pact verification results ? and I can?t see where to hook in the command to disable webmock

```WebMock::NetConnectNotAllowedError: Real HTTP connections are disabled. Unregistered request: GET https://site.pact.dius.com.au/pacts/provider/.../diff/previous-distinct with headers {'Accept'=>'*/*', 'Accept-Encoding'=>'gzip;q=1.0,deflate;q=0.6,identity;q=0.3', 'User-Agent'=>'Ruby'}```

this is using the `publish_verification_results` setting in the `Pact.service_provider` declaration to trigger uploading the results

I have it in my pact_helper.rb file

I tried that and couldn?t get it working

Instead I added this to my `rails_helper.rb`: ```WebMock.disable_net_connect!(allow_localhost: true, allow: "http://site.pact.dius.com.au")```

How have you set it up @qatrera?

Well I use and old version of `Typhoeus` instead of `WebMock` so maybe what works for me it's not transferable to you. And we also have the pact broker hosted in house

I would have disabled WebMock for that domain the same way.

has joined #pact-ruby

@bethskurrie @matt.fellows https://github.com/pact-foundation/pact-ruby-standalone this is to replace all the gems and use one for everything both on provider and consumer side

has joined #pact-ruby

Ruby folks here Do you people use Semantic versioning to version pacts or pass commit hash as input for pacts ?

have a read of https://github.com/pact-foundation/pact_broker/wiki/Pacticipant-version-numbers if you haven?t already

This might also be of use: https://docs.pact.io/getting_started/versioning_in_the_pact_broker

I got this but the problem comes here is ho do we do it through the code like our semantic versioning only sits in the releases

@bethskurrie any ideas?

Sits?

If you know or can work out your sha at release time, I'd recommend using a git sha. If you use your semantic version, you'll tend to have multiple commits, and hence, builds and pacts per minor or patch version, and that means you're overwriting the previous pact version with that semantic version. You can do it, but things tend to work better when your data is immutable (ie one pact version per consumer version number).

You eliminate race conditions, and you get a more fine grained "matrix" of test results.

@bheemreddy181 when you say "sits in the release" do you mean you don't version things until you release? What's the process you use to version now?

Ya that?s true we don?t tag the app until we plan to release

You don?t have to give your published app version the exact tag you push to the broker (although it probably is more convenient).

@bethskurrie Quick Question whenever i use the broker - do i need to give pact_uri in both these places ? https://github.com/pact-foundation/pact-ruby-e2e-example/blob/c54dd9da813c58261b9a83e6249b1d62920fb72b/provider/spec/pact_helper.rb#L11 , https://github.com/pact-foundation/pact-ruby-e2e-example/blob/c54dd9da813c58261b9a83e6249b1d62920fb72b/provider/Rakefile#L27 ?

They're two different ways of doing the same thing.

Typically, you would do it the first way.

You can add extra verification tasks if you want to do custom verifications (eg verify a pact on your local machine)

You should give the pact helper configuration the broker details though, and it will fetch all the pacts from the broker. You don't have to use the specific urls.

Little confused, can you please elaborate

For the last point

@bethskurrie Here is another quick question - On this verification task, the version belongs to Consumer or Provider ? https://github.com/pact-foundation/pact-ruby-e2e-example/blob/c54dd9da813c58261b9a83e6249b1d62920fb72b/provider/Rakefile#L27

consumer

typically you'd use `pacts/provider/Bar/consumer/Foo/latest` though

you don't want to fix it to a particular version.

@bheemreddy181 this is the recommended configuration https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#fetching-pacts-from-a-pact-broker

when you say consumer how will the provider know which version he has to verify with @matt.fellows

the latest version should be used generally

you dont' fix it to a particular version. But, it is recommended that you fix it to a particular tag. eg "the latest dev pact" or "the latest prod pact"

Beth, is there a way in Ruby to just point at the broker with the provider details and get all of the consumers for a given provider?

the docs are in the link I sent above, or you can use `pacts/provider/Bar/consumer/Foo/latest/TAG_NAME`

yes

the docs in the link I posted above describe it

`#fetching-pacts-from-a-pact-broker`

oh sorry, skipped over that link

what does this mean ? ``` But, it is recommended that you fix it to a particular tag. eg "the latest dev pact" or "the latest prod pact"```

@bheemreddy181 I?d 100% use the https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#fetching-pacts-from-a-pact-broker strategy. It?s by far the most straightforward way for anything but the most simplest pact setups

have you used tags yet?

how will the provider know which version of my consumer is in

it uses the latest one

it doesn't care which version it is. it's just whatever the latest version of the pact is.

what if I want to use the commit hash?

it still uses the most recently published version

it's time based, not semantic

Does the above link still applies when i use commit hash ?

yes

Sorry if i am bothering you , i am trying to understand

``` honours_pacts_from_pact_broker do pact_broker_base_url 'http://...' end ```

How about as a rake task ? So that i can publish the results back

it publishes the results back automatically

actually, here are the docs

Please just read this page top to bottom, and then ask any questions you have.

Sure

has joined #pact-ruby

Hello there! I?m wondering if anyone?s using the standalone mock server as explained here? # standalone true

oops

here: https://github.com/pact-foundation/pact-ruby/wiki/RAQ#how-can-i-run-a-standalone-mock-server

:cricket:

This gets used a lot: https://github.com/pact-foundation/pact-ruby-standalone/releases but might not be what you?re after

We just use the git commit sha

is there a way i can use let blocks in pact provider states

@bethskurrie any idea ^

Here are few more questions : 1. Parma type validations on the API - does required or optional param should happen on the controller of the API - As we are mocking everything underlying logic of the controller what if if you param validations happens inside the some where with in the logic ?

Parameters* type validations

@matt.fellows ^

Oh, I had a link on that, went to look for it, got distracted.

@bethskurrie ^ similar to rspec https://relishapp.com/rspec/rspec-core/docs/helper-methods/let-and-let

No, but you can include helper methods

which helper methods you mean ?

```Q: I felt like PACT does give a provision of breaking contracts if none of your consumers is using few fields from your API or not bothered if the type is changed? yet all? Is that true? which might also lead to having just one version of the API unless needed? A: Exactly true, consumer contracts remove the need to version an API. API versioning is a "workaround" because you don't want to break backwards compatibility, and you don't know who is using your API. It's the best technique for a public API. When you know all your consumers, and you know exactly what they're using, you know the impact of change, and hence, you should be able to avoid versioning the API. Q: Can I do a functional test of my provider using PACT? A: You can, but you are not supposed to do that . Give a read from here: https://docs.pact.io/best_practices/consumer/contract_tests_not_functional_tests Q: Do I need to write contracts tests for the complete response of an API? A: No, That's a bad way of writing tests - From Consumer perspective you have to write tests for the fields only you worry about. Q: Why can't I generate a schema from Provider Side and use that as a reference for all future changes? A: Yes you can do that but when your API is public like PUBLIC API and you don't know how your consumers are using your API, else you are tightly coupling your changes to all the fields which are not generally required. Again, you are duplicating the functional tests with your contract tests from the provider side. which should always be consumer driven. Q: Can I generate contract file from Swagger? A: Give a read: https://docs.pact.io/faq#can-i-generate-my-pact-file-from-something-like-swagger Q: Oh Wow, Who will write the Contracts test - which will generate the contract files? A: Its the consumer duty to write the tests pretty much to specific to the fields what consumer bothers rather than writing for complete body validations. on the provider side just that you have to maintain the states of mocks - like Skip authentication so on. Q: Oh This looks like VCR recording, how is it different from that? A: Give a read: https://docs.pact.io/faq/convinceme#but-i-already-have-a-local-mock-server-e-g-vcr-mockserver Q: who will own it, like who will own the contract files? A: Consumer again, each provider can have n - number of consumers and each one might use a different one. So it's the consumer duty from what he uses to write the contract from a consumer perspective. Q: What is PACT_BROKER? Is this a hosted service which should live somewhere and someone should own it? A: Yes this is a hosted service and should be owned by someone. Currently, PACT developers provide a free hosted version of pact_borker with security attached if we want we use that. Else we can have a dedicated host where this service can live, they have a docker image which we can use to host this on our side and do upgrades. Sample free host here: https://pactflow.io/register/ Q: where will it run? A: Our Aim is to run this each Pull-request on Travis/ANY CI TOOL. Q: when will it run? A: On Each Pull Request Q: Do we have to write some code on the provider side? A: StuB the internal functionality of the provider, because of FUNCTIONAL TESTING is the job of the provider to test than tested using consumer driver contracts. Make sure, if you stub, that you don't stub the code that actually parses the request and pulls the expected data out because otherwise, the consumer could be sending absolute rubbish, and the pact:verify won't fail because that code won't get executed ```

@bethskurrie @matt.fellows are you people inline with the above questions and relative answers , correct me if i am wrong.

Typically on Consumer PR perspective

On Provider PR

Same as these too

Looks good to me. I would say a flat "no" to "can I use pact for functional tests" though.

Perfect :+1::skin-tone-3:

@luchillo17 little notes what I prepared would help you understand more ^

in the first diagram I would say the deploy arrow in the provider is not needed right? I also don't get the Skip section in the second diagram

Maybe it's worthy to extend the FAQ (https://docs.pact.io/faq) with these Q&As. I found some of them really valid and eye-opening for Pact newbies :slightly_smiling_face:

Really nice diagram btw :slightly_smiling_face: I still need to find time to dig into the webhooks part to finally integrate the first contract tests in our company :slightly_smiling_face:

here is a question : i have a rake task in my rake file as this : ```Pact::VerificationTask.new(:portfolio_using_remote_broker) do | pact | pact.uri "#{REMOTE_PACT_BROKER_BASE_URL}/pacts/provider/PGS/consumer/Portfolio/latest", :pact_helper => 'spec/service_consumers/pact_helper.rb' end```

whereas in my pact helper : ```require_relative './provider_states' Pact.configuration.reports_dir = "./provider/reports" Pact.configure do | config | config.diff_formatter = :embedded end Pact.service_provider "PGS" do app_version '1.0.0' publish_verification_results !!ENV['PUBLISH_VERIFICATIONS_RESULTS'] honours_pact_with 'Portfolio' do # This example points to a local file, however, on a real project with a continuous # integration box, you would use a [Pact Broker](https://github.com/bethesque/pact_broker) or publish your pacts as artifacts, # and point the pact_uri to the pact published by the last successful build. pact_uri 'https://remote:endpoint', {username: '...', password: '...'} end end```

whenever i do run rake task ``` be rake pact:verify:portfolio_using_remote_broker ```

i get 401 unauthorized

i did pass the usernames and passwords as requested in the pact_helper

it does work when i do append the pact_uri with /pacts/provider/PGS/consumer/Portfolio/latest on the pact help and just run the pact:verify task but that doesn't publish the results

pact_helper*

Skip section in the second diagram comes into picture when you publish the same contract every time

You can post your questions here I will post my answers

@bethskurrie ^ some help on the above stuff ?

@bethskurrie Can i get some help on the above stuff

> i get 401 unauthorized

from what?

the broker? the application under verification?

from broker

have you tried curling the URL and providing the basic auth username and password?

Is it a local broker you?re hosting or our hosted version?

Hosted version

have you tried curling the URL and providing the basic auth username and password?

curl PACT_URL -u username:password

that worked ```{"_links":{"self":{"href":"https://bheemreddy.pact.dius.com.au","title":"Index","templated":false},"pb:publish-pact":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/consumer/{consumer}/version/{consumerApplicationVersion}","title":"Publish a pact","templated":true},"pb:latest-pact-versions":{"href":"https://bheemreddy.pact.dius.com.au/pacts/latest","title":"Latest pact versions","templated":false},"pb:tagged-pact-versions":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/consumer/{consumer}/tag/{tag}","title":"All versions of a pact for a given consumer, provider and consumer version tag","templated":false},"pb:pacticipants":{"href":"https://bheemreddy.pact.dius.com.au/pacticipants","title":"Pacticipants","templated":false},"pb:latest-provider-pacts":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/latest","title":"Latest pacts by provider","templated":true},"pb:latest-provider-pacts-with-tag":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/latest/{tag}","title":"Latest pacts for provider with the specified tag","templated":true},"pb:provider-pacts-with-tag":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/tag/{tag}","title":"All pact versions for the provider with the specified consumer version tag","templated":true},"pb:provider-pacts":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}","title":"All pact versions for the specified provider","templated":true},"pb:latest-version":{"href":"https://bheemreddy.pact.dius.com.au/pacticipants/{pacticipant}/latest-version","title":"Latest pacticipant version","templated":true},"pb:latest-tagged-version":{"href":"https://bheemreddy.pact.dius.com.au/pacticipants/{pacticipant}/latest-version/{tag}","title":"Latest pacticipant version with the specified tag","templated":true},"pb:webhooks":{"href":"https://bheemreddy.pact.dius.com.au/webhooks","title":"Webhooks","templated":false},"pb:integrations":{"href":"https://bheemreddy.pact.dius.com.au/integrations","title":"Integrations","templated":false},"beta:pending-provider-pacts":{"href":"https://bheemreddy.pact.dius.com.au/pacts/provider/{provider}/pending","title":"Pending pact versions for the specified provider","templated":true},"curies":[{"name":"pb","href":"https://bheemreddy.pact.dius.com.au/doc/{rel}?context=index","templated":true},{"name":"beta","href":"https://bheemreddy.pact.dius.com.au/doc/{rel}?context=index","templated":true}]}}```

well, now you know you have the right creds at least

you've either configured it incorrectly, or there's a bug in the ruby code.

help me understand one thing here

when i have my own rake task the credentails are shared from the pact helper ?

this is my rake task

```Pact::VerificationTask.new(:portfolio_using_remote_broker) do | pact | pact.uri "#{REMOTE_PACT_BROKER_BASE_URL}/pacts/provider/PGS/consumer/Portfolio/latest", :pact_helper => 'spec/service_consumers/pact_helper.rb' end```

where i am not using any creds

but i my pact helper i have these

the rake task is used when you want to override the default behaviour in the pact_helper.rb

```require_relative './provider_states' Pact.configuration.reports_dir = "./provider/reports" Pact.configure do | config | config.diff_formatter = :embedded end Pact.service_provider "PGS" do app_version '1.0.0' publish_verification_results !!ENV['PUBLISH_VERIFICATIONS_RESULTS'] honours_pact_with 'Portfolio' do # This example points to a local file, however, on a real project with a continuous # integration box, you would use a [Pact Broker](https://github.com/bethesque/pact_broker) or publish your pacts as artifacts, # and point the pact_uri to the pact published by the last successful build. pact_uri 'https://remote:endpoint', {username: '...', password: '...'} end end```

you'll need to give the task the credentials, because you're overriding where the pact comes from

how can i pass the credentials in my rake task then ?

but in that cases does my pact helper still needs pact_uri and creds ?

`With basic auth, set the environment variables PACT_BROKER_USERNAME and PACT_BROKER_PASSWORD, or use the basic auth URL format, http://username:password@pactbroker.yourdomain/.... `

why are you making a custom verification task?

So that I can add that to other rake task which I run as part of my CI build

why do they need to be different?

When I run for pact helper the verification results aren?t published

you just configure that

`publish_verification_results !!ENV['PUBLISH_VERIFICATIONS_RESULTS']`

on your CI server, you set that env var to true

when you run it locally, it's not set, so the results dont' get published

you shouldn't run a different task.

it's the same task, just configurable.

Makes sense

Is there something that needs improving in these docs, or did you just miss that bit?

I felt they were not that clear for a new comer

Are you unfamiliar with Ruby itself?

Not really but I am not a developer though

Ah, I see.

Thanks for inputs

I've updated that section - is it clearer? https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#publishing-verification-results-to-a-pact-broker

Clear now

Thanks

:thumbsup::skin-tone-3:

One more thing when I say provider/consumer/latest does that verify all the pacts which say tags prod and Dev ?

If I had provided that tags as prod , Dev

the latest pact for each of those tags

@bheemreddy181 if you make a module that has helper method for creating or stubbing your data, you can include that module directly into the provider states code. I think it's something like `Pact.configuration.include YourHelperModule`

You want to push as much logic into the helper code, otherwise your provider state set up code will get very bloated

I personally like a "builder" pattern. Here's the provider state code for the pact between the Pact Broker and the Pact Broker Client (meta!) https://github.com/pact-foundation/pact_broker/blob/master/spec/service_consumers/provider_states_for_pact_broker_client.rb

has joined #pact-ruby

has joined #pact-ruby

Can some one explain how can i do the similar setup in ruby

```... parameters { string(name: 'pactConsumerTags', defaultValue: 'master') } stage ('Build') { ... sh "./mvnw clean verify ... -Dpactbroker.tags=${params.pactConsumerTags}" ... } }```

Does the consumer - triggered provider verification job should be different from provider driven verification should be different from each other ?

Same way inputs on this @bethskurrie ^

@matt.fellows ^

I don?t see a way to set a default value in there do I need to write my own helpers for the same

You need to know how to write ruby

I think you'll have to get some help from your own developers.

I am getting confused a bit here

Yes, I can see that.

Help me understand two use cases here

I think what you're asking me to help you with is beyond the scope of what I can achieve over slack.

Ok, I can help you with usecases, but you seemed to be asking for ruby coding help.

Why does provider need to same branch name as consumer when sends the branch names as a tag

it doesn't

some people like to make a matching branch name in the provider to implement the required features, but it's not mandatory.

In that case too when the branch_name is considered I should fallback to master when not present

And when consumertagname is present then I should fetch that else always master

hm.

Yes

Consumer_tag_versions {Consumertagname || master , travis[?branch_name]}

if you're not using provider branches for feature development, you can ignore that part though.

This should be something like this

I'm pretty sure it states that that is an advanced use case

Yes does above tags makes sense

? The way it should be set like the same

I don't know where you're getting Consumertagname from

As part of webhook Params

When triggered by consumer job

yes, but how are you getting it into the ruby process?

That I should figure it out don?t know as of now

I would use environment variables

Cool that makes sense

``` consumer_version_tag = ENV['CONSUMER_VERSION_TAG'] || ENV['TRAVIS_BRANCH'] consumer_version_tags [{name: consumer_version_tag, fallback: 'master'}] ```

Perfect cool :sunglasses: thanks Beth this is a great help

There. That will take the consumer tag name if it exists, or the travis branch, or fall back to master

I am so dumb and eating up all your time

:clap::skin-tone-3::clap::skin-tone-3::clap::skin-tone-3::clap::skin-tone-3: Appreciate your help

You're obviously trying to do something in a language with which you're not familiar. I admire your perseverance, but I think you need to find someone in your company to help you out with the ruby stuff.

I can't be a ruby help desk to everyone!

I understand that makes sense. Appreciate your help on the same I think I am clear as of now

has joined #pact-ruby

has joined #pact-ruby

Hi everyone, I am not sure if this is correct channel to ask this question, I am trying to run some provider tests using docker compose -> image: dius/pact-provider-verifier-docker My docker compose file looks like this: ``` pactverifier: image: dius/pact-provider-verifier-docker links: - app:app environment: - pact_urls=https://myteamtbroker.pact.dius.com.au/pacts/provider/providername/consumer/consumername/latest - provider_base_url=http://app:8000``` However, I am getting this error: ```pactverifier_1 | /usr/local/lib/ruby/2.6.0/rubygems.rb:283:in `find_spec_for_exe': Could not find 'bundler' (2.0.1) required by your /app/Gemfile.lock. (Gem::GemNotFoundException) pactverifier_1 | To update to the latest version installed on your system, run `bundle update --bundler`. pactverifier_1 | To install the missing version, run `gem install bundler:2.0.1` pactverifier_1 | from /usr/local/lib/ruby/2.6.0/rubygems.rb:302:in `activate_bin_path' pactverifier_1 | from /usr/local/bundle/bin/bundle:23:in `<main>'``` I am not familiar with ruby, am I missing something?

Hmm that doesn?t look like it should happen.

What?s your provider code written in, Aniela?

Javascipt

Is there a reason you?re not using the JS based verification?

I tired, but I got issues running that inside a docker container

@aniela.cole there's a newer one, let me point you to it

What issues?

It should work just fine in Docker (that?s almost always how I run them)

@aniela.cole try this: https://github.com/pact-foundation/pact-ruby-cli#verify-pacts

Just seen Matt's chat though - yes the pact-js verifier should run on docker

yes, I am just running it now to get the issues to send to him. Thanks!

```/bin/sh: ./standalone/linux-x64-1.68.0/bin/pact-provider-verifier: not found at ChildProcess.<anonymous> (node_modules/@pact-foundation/pact-node/src/verifier.ts:161:55)```

Are you using Alpine?

Have you seen https://github.com/pact-foundation/pact-js/issues/296

yes, I have, I am already doing that

hang on, I did not have this `&& apk add glibc-2.29-r0.apk ` let me add it

:+1:

I am still getting the same error ```/bin/sh: ./standalone/linux-x64-1.68.0/bin/pact-provider-verifier: not found lerna ERR! lerna ERR! at ChildProcess.<anonymous> (node_modules/@pact-foundation/pact-node/src/verifier.ts:161:55)```

Can you run the example project linked in that issue? Something about the environment must be wrong

has joined #pact-ruby

Matt, my issue was caused by having this volume on my docker-compose file ``` volumes: - ./:/app ``` and having the node_modules locally.

once we removed the volume, it worked fine

ah, because the node_modules were OS specific?

Thanks for taking the time to respond - most people never come back to provide that feedback :thumbsup:

yes, I believe it was causing the problem. Thank you guys for helping out!

has joined #pact-ruby

has joined #pact-ruby

@krzysztof.stepniak has left the channel

has joined #pact-ruby

has joined #pact-ruby

Is there a way to use the mocking service without writing to the pact file? The idea is to use a feature flag in order to be able to write consumer code, test it and even merge it to master, but without changing the pact in the broker and breaking the provider build.

has joined #pact-ruby

has joined #pact-ruby

:wave: Hi! A question, how can I define the provider base url for contract verification from code (as opposed to the CLI). I know I can do it with `pact-js` in the Verifier options. Our use case: we have two contracts against different endpoints implemented in the same code base, but the endpoints for each contract are only available under different subdomains. So, contract `x` needs to be verified against `x.localhost` and contract `y` against `y.localhost`. Otherwise the endpoints will be reported as not found.

Does the Ruby DSL not support this? Given that the pact JS, Go and a number of others depend on it that would be surprising

has joined #pact-ruby

If you're using a native Ruby verification, then you're running the verification against a Rack app, and there is no base url.

What is your code for running the verification currently?

It?s a rails API, a monolithic API. A single codebase/process runs multiple providers. Each provider runs in a different subdomain in production. There is a check in place inside the Rails app to expose some routes in one subdomain and other routes in another subdomain. Rails routing constraints are used for that. So, I thought we could configure the `Pact.service_provider` for each provider to attack a different domain `provider_a.localhost`, `provider_b.localhost`, etc.

It seems that I can do that with `pact-js` using `providerBaseUrl` in the `Verifier` options

An alternative approach I?m thinking of would be to run the `rake pact:verify` in a specific Rails environment `contract-verification` where we configure all the providers without subdomain

Do you know if we can pass the rails environment to the `pact:verify` task?

@enrique you can use rack middleware to modify the request.

By default, with rack test, the domain is http://example.org

See https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#modifying-the-request-with-live-data

In Ruby, you don't use a running server. You test directly against the rack app @matt.fellows

Ah yes. I was just saying I'd like to be able to do this in JS and Go with Ron just today

Yup, that'll be a really nice feature.

Is there a Rack equiv for other languages though? I got the impression that Ruby was pretty unique in having such a standardised interface.

don?t feel so special Beth :slightly_smiling_face:

Yes, it?s common

This is very common in Go

Less so in JS, but the types are types and it works exactly as you?d expect

Node JS has an HTTP interface that most libraries (e.g. Express) use. Go has the same sort of thing, but goes one deeper and has an in-built HTTP testing package (i.e. you can do all of the things without pulling in a single external dependency)

Awesome sauce.

Oh, yes that would work. We also finally come out with a different approach. But, now, as you mention, the domain is set to `http://example.org`.

We can work around that, but it feels wrong. Guess we need to configure rack test somehow

Does `pact-ruby` expose any configuration option to change the domain of `rack-test`? That would be the ideal solution. It seems that we cannot do it outside PACT because we need access to the app used by PACT.

It gets set by rack test

You can change it with middleware, as described above.

There's nothing wrong with that approach.

It's what I would do.

Hi! I?m trying to publish the verification results to the broker as described in the docs https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#publishing-verification-results-to-a-pact-broker but for some reason I keep getting the following error: ``` Error returned from verification results publication 400 {"errors"=>{"provider_version"=>["can't be blank", "Version number '' cannot be parsed to a version number. The expected format (unless this configuration has been overridden) is a semantic version. eg. 1.3.0 or 2.0.4.rc1"]}} (Pact::Provider::VerificationResults::PublicationError) ``` While debugging I see that the `provider_application_version` attribute of my `Pact::Provider::VerificationResults::VerificationResult` instance is not blank at all. Is it a naming thing? The broker expects `provider_version` but the Ruby library sends `provider_application_version` maybe? Any ideas?

has joined #pact-ruby

has joined #pact-ruby

Hey @bethskurrie, looks like the travis build of the pact ruby gem is broken? On every new PR it seems.

:eyes: It could be cause by some changes committed by mistake with the release script? https://github.com/pact-foundation/pact-ruby/commit/7371e2f9fc66491481288a2902630419b3f1debc#diff-249018b712e330bd9b05de4e31a5477aR5-R7

I've just reverted those gemfiles. I accidentally committed changes from my local testing.

Thanks! I rebased my PR on master and it looks good now.

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

I have a question that I suspect may have the answer (a) that is not what Pact is for or (b) that has already been answered ? (and possibly also (c) why would you want to do that?). Sorry if this is the the case. For an API (pact provider, I think is the correct terminology) I would like to have a wrapper library that deals with all the http calls. This wrapper library would then be used by the client application instead of making all the API calls explicitly. I know that I can create the contract between the wrapper library and the API but, in terms of checking the contract during deployment, I want to do this in the client application. Is there a correct way to set this up? Should I (and is it possible to) create a contract between the client application and the wrapper library? I have already done this (some time ago) and the solution I came up with is that the client application copies the contract from the wrapper library and submits that as its own, but this doesn?t feel right.

I?ve posted this in this channel because I?m mainly working in ruby & rails but it occurs to me that the question may be more general.

I?ve posted it again in #general but I don?t appear to have the option to delete. :disappointed:

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Does anybody have an idea what's going on with the assets on https://github.com/pact-foundation/pact-ruby-standalone/releases/? We cannot download them anymore. This also breaks our pip install pact-python, as that will try to download the binaries there.

What?s the issue you?re having? I can download them just fine (I?m on a train tethering with my wifi hotspot)

Just downloaded 1.70.2 for linux

Maybe it?s your corporate network / proxy / firewall?

@aniela.cole has left the channel

Issues seemed to be a DNS issue indeed. Using Google DNS servers it just works fine.

@matt.fellows @bethskurrie are dB inserts bad as part of pact workflow on the provider side ?

Depends.

Personally, I've used mostly used pact for microservices, and for our pact verification tests, we usually use something like sqlite, so it's really easy just to use the DB.

Typically I am talking more in regards to post endpoint which inserts data into dB I am mocking all the validations that end point does but directly inserting that data into dB because that what my presenter uses to render response back

If it's some heavy weight DB that it would be a pain to use in what should be a fast test, then mocking might be more convenient for you.

My tests also does cleanup so it very light weight

Why would you mock all the validations?

wouldn't that make the endpoint behave differently to real life?

I don?t mock validations on required fields but mock other validations which are supposed to be tested by the team who owns it

If I don?t mock then isn?t that a functional test itself

I don't quite understand, but sounds dicey

functional isn't bad.

if it's just a side effect of testing the contract.

it's when you try and use a contract to do functional that it becomes a bad idea.

If the validation changes the way the endpoint behaves, then it's part of the contract.

Eg. if I was trying to test how the endpoint behaved when I put in invalid data, it would not be helpful to return a 200 OK!

Let me read this and get back

Thanks for sharing

Something that might help is to ask, if I stubbed out this component, what other tests do I have to make sure that it will all work together correctly (trying not to rely too much on e2e tests). What would be the consequence of stubbing something out and missing some coverage vs the consequences keeping it in?

I mean I have enough functional tests on the provider side which tests these validations

From my understanding I am making sure the return types on my presenter and fields aren?t dropped or required fields aren?t changed etc etc

is there a pact test that checks how the endpoint responds if the validation fails?

From my contact tests

Nope

I'd recommend asking the consumer team to include that.

But in that case each request values has different validations so they have to add all that combinations?

With Pact, you *don't* want to try and test the business rules that cause the validation to fail, but you want to pick a stable scenario that will cause the validation to fail so that the consumer knows that that response looks like (assuming all validation failure responses have the same pattern)

You could, of course, cause the validation to fail by modifying something in the provider state, rather than expecting some type of data from the consumer.

What all should sit in provider state then in general?

I'm sorry, the question doesn't make sense to me.

Can you rephrase it?

When ever we write provider states in provider what all mocks should generally sit there

I personally mock downstream APIs, and anything that's too hard/slow to run as part of the verification test.

Aren?t pact tests become duplicates of the provider functional tests if we write them at both places ?

Did you read the link I sent about functional vs contract tests?

You mean if there are any third party API?s underneath my provider api ?

yes

I will read that for sure

ok, read and then come back and we can discuss if it's still unclear.

no point repeating all that content.

Sure Beth

Cool it makes more sense for me

Awesome

Here are some use cases I am worried about date and time dependencies

Like say if I can?t create a payment after 10:30 pm cst but if I am running a test after that it would fail ?

But before 10:30 I will get a 200 ok this will not happen in production because the configurations or jobs which run are restricted to run a hour before

But let me go back and see all use cases and come back with a list if there are any

I would mock the time, not the validation

Ensuring you always get current time from a mockable location is always a good thing for testing

Though, now I think about it, your approach makes sense.

As long as you have a test that shows how it works when it succeeds, and when it fails.

Ya that?s where I was mocking some of the underlying validations but I will go back and see which all use-case needs these

@bethskurrie I have another use case where incoming request should not have date in the past in this you mean me mock the incoming time to current all the time and few are cutoff based requests like said above payment can be created given a :bank: bank is fixed cutoff

I mean tomorrow if bank is extending the cutoff shouldn?t break the contracts correct those are not breaking contracts too

Is this a unique response? ie. is there anything different about the shape of the response body that comes back when the time cutoff has occurred, compared to other validation errors?

If request comes in post cutoff then it?s a 422

Validation failure

What language are you validating in?

Ruby

Is the validation error different in shape to other validation errors?

Nope all fall in the same bucket just message is different

I personally wouldn't bother testing it then.

That?s where I was mocking the validations for these so you mean these are valid mocks ?

All you need is one example of "this is what the response looks like when validation fails". Doing that multiple times is just overhead.

Which turns out to be a functional test correct ?

Hm. It's not so much functional as redundant.

functional tests check for side effects, which this is not doing.

it's just unnecessary.

You're creating a unnecessary maintenance burden.

Can you give me a short example of side effects in general

From the docs, "a functional test for the provider would ensure that when a given request was made, that an Order with the correct attributes was actually persisted to the underlying datastore"

Functional tests check "does it do the right thing with what I've sent". Contract tests just check "does it respond in the expected way".

I'd say that functional tests were a superset of contract verification tests -> they test the response is as expected AND that the right thing has been done with the data.

Makes sense

Hey Beth, i can?t mock future dates all the time as this forces me to update my contract my consumer all the time , if my consumer doesn?t change we will keep verifying with old contract which has the old date as input

which throws a error like below ```{"errors":[{"error":"validation_failed","reasons":{"effective_on":["Effective date in past."]}}]}```

So, you can either use rack middleware to update the date (look in the "verifying pacts" docs on the Ruby wiki) or you can stub "now".

something like this https://github.com/pact-foundation/pact-ruby/wiki/Verifying-pacts#modifying-the-request-with-live-data

Yes

Is there a way we can run rack middleware for only pact tests

Yes

Your can set up the rack app manually. It's documented on that same page.

It's in the example section that you sent me.

Beth , thanks for all your inputs i was able to set up rack middleware for one use case

@khoatran

has joined #pact-ruby

@khoatran now I've refreshed my memory, I've remembered the DSL for message pact is not done in Ruby yet. The underlying logic is done though, because it's being used by pact-js.

Which side would you be interested in using? Consumer or provider, or both?

Oops, now I see the #pact-message-ruby channel that I'd forgotten about! Moving over there.

has joined #pact-ruby

Just to verify, are we not supposed to test the data semantics using contract tests @matt.fellows @bethskurrie ?

I would agree with Tom?s article and that statement, yes

Well I am surprised here whether we have to mock the provider side functionality or not now

As I heard last time from you is that we shouldn?t mock other than any other service dependencies( like if I have third party API?s ) using contract testing

@matt.fellows@matt.fellows@matt.fellows ^

I don't know what is meant by data semantics here.

If you're talking about the difference between a validation error response, and a success response, we want to cover that.

If we're talking about the differences between all 20 different enumerations of some particular field, then we don't want to cover that.

There might be a semantic difference between each value, but no structural difference to the response.

Can you provide me with a concrete example that you're unsure about?

The example used is a credit card

do we validate all of the permutations of valid, invalid (and so on), or do we just care about the unique set of responses we can illicit from the provider

from a consumers? perspective, we just care about the possible responses. The meanings of the card numbers are something the provider is interested

Nice example

Ah I agree with this as well

Well from what I felt is semantics is how you are validating the data and inserting the data on to the database where are as syntax is how you are presenting the data back in your response

What you want to test is what a validation failure response looks like. To get that, the easiest way is to give it invalid data, but it's not the validation logic that you really care about.

You could mock some internals of the provider to force it to return a validation failure, but it may be more effort than its worth.

:+1::skin-tone-3:

Can I get some help on the above ? ^

has joined #pact-ruby

Oh, I thought the mock process in Ruby was purely Rack based, so it didn?t need to spin up an actual port at all.

But admittedly, it?s been a while since I used the Ruby Pact DSL (like, 4 years :stuck_out_tongue: )

@bethskurrie ^ In case you missed i am open to create a request on the ruby implementation

thats not true

@bheemreddy181 if you don't give it a port, it should pick one for you

You'll need to then use a method on the dsl base object to get the base url that it chose.

you mean to say for mocking the response from the provider i need to get the base url it chose ?

Yes

I think it's mock_service_base_url

No, I mean, when you make the real request

You need to know the base url of the mock service to send it to

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

@bheemreddy181 Hi Bheem, I am trying to setup PACT in ruby and trying to run basic example - Animal Service and Zoo App just for understanding purposes. I am able to generate PACT file. After that, when i try command rake pact:verify, nothing happens. It doesn't give me either failure or success messages. I tried pact:publish to see if file gets published to localhost:9292(its by default pointing to this from PACT repo) but getting error saying connection refused. Looks like you got success with setup. Can you please help?

@bethskurrie @matt.fellows I am pretty new to Pact. I am good until generation of Pact file(atleast was able to generate for the animal-zoo example). After that, rake pact:verify is doing nothing. am i missing some steps here?

I could install pact-provider-verifier gem and ran "pact-provider-verifier zoo_app-animal_service.json --provider-base-url http://localhost:9292" command, got all 3 interactions failed:

C:\Automation\PACT\pact-ruby-master\example\animal-service\spec\pacts>pact-provider-verifier zoo_app-animal_service.json --provider-base-url http://localhost:9292 INFO: Reading pact at zoo_app-animal_service.json Verifying a pact between Zoo App and Animal Service Given there is an alligator named Mary a request for an alligator with GET /alligators/Mary returns a response which WARN: Skipping set up for provider state 'there is an alligator named Mary' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. has status code 200 (FAILED - 1) WARN: Skipping set up for provider state 'there is an alligator named Mary' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. has a matching body (FAILED - 2) includes headers WARN: Skipping set up for provider state 'there is an alligator named Mary' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. "Content-Type" which equals "application/json;charset=utf-8" (FAILED - 3) Given there is not an alligator named Mary a request for an alligator with GET /alligators/Mary returns a response which WARN: Skipping set up for provider state 'there is not an alligator named Mary' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. has status code 404 (FAILED - 4) Given an error occurs retrieving an alligator a request for an alligator with GET /alligators/Mary returns a response which WARN: Skipping set up for provider state 'an error occurs retrieving an alligator' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. has status code 500 (FAILED - 5) WARN: Skipping set up for provider state 'an error occurs retrieving an alligator' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. has a matching body (FAILED - 6) includes headers WARN: Skipping set up for provider state 'an error occurs retrieving an alligator' for consumer 'Zoo App' as there is no --provider-states-setup-url specified. "Content-Type" which equals "application/json;charset=utf-8" (FAILED - 7) Failures: 1) Verifying a pact between Zoo App and Animal Service Given there is an alligator named Mary a request for an alligator with GET /alligators/Mary returns a response wh ich has status code 200 Failure/Error: replay_interaction interaction, options[:request_customizer] Errno::ECONNREFUSED: Failed to open TCP connection to localhost:9292 (No connection could be made because the target machine actively refused it. - connect(2) for "localhost" port 9292 ) # C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-provider-verifier-1.25.2/bin/pact-provider-verifier:3:in `<top (required)>' # C:/Ruby25-x64/bin/pact-provider-verifier:23:in `load' # C:/Ruby25-x64/bin/pact-provider-verifier:23:in `<main>' # ------------------ # --- Caused by: --- # Errno::ECONNREFUSED: # No connection could be made because the target machine actively refused it. - connect(2) for "localhost" port 9292 # C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-provider-verifier-1.25.2/bin/pact-provider-verifier:3:in `<top (required)>'

Which directory are you running `bundle exec rack pact:verify` from @shilpa3006?

That's not going to work unless you've manually started the Animal Service on port 9292.

I'd stick with the native Ruby pact:verify.

Oh, I can see from your example, you're in the right directory.

Are you familiar with developing in Ruby at all @shilpa3006?

Ah - I see you're running the examples on windows.

I think that's likely to be the cause. I'm not familiar with running Ruby on Windows I'm afraid.

Is there anyone else in your company you could ask who might know about Ruby+Windows?

Yes Beth. I am familiar with writing BDD scripts in Ruby. I did so for most of our applications for test automation purposes with Ruby, Cucumber framework. But, I haven't developed applications itself with Ruby.

hmm. now i am curious. which OS is recommended for this pact framework? Also, I don't have Rails. It would be better if there would be any prerequisites w.r.t OS in documentation.

I didn't mention rails.

Not all Ruby is Rails!

The native Ruby pact verify will manage the lifecycle of the app for you - you don't need to start and stop it and make extra endpoints for setting up the provider states.

It's more complex to use the pact provider verifier.

It'll be something to do with Windows. If you can reproduce the issue with a Windows ci system as I mentioned in the issue, I may be able to help you.

I wish pact:verify works for me. How can i troubleshoot this issue? I don't see any logging for this as well.

I guess you don't have any other machine apart from windows?

Nobody else has ever reported this issue before.

not at the moment. can you pls send me that link where i can reproduce. I saw earlier. Somehow i lost it.

It's in the issue.

Where you reported it.

ok

looks like i will not be able to do this. Its against our comp policy. I own a mac, i guess i can try there but eventually i need it at my work :disappointed:

What is against your company policy?

https://github.com/pact-foundation/pact-ruby-e2e-example better to start with this example

I will try this and keep you all posted. Thanks @bheemreddy181

Thanks @bheemreddy181! I'd forgotten to suggest that.

I can help the beginners now :grimacing:

@bheemreddy181 I have tried this and still pact:verify doesn't work.

What?s the issue

it does't do anything @bheemreddy181

C:\Automation\PACT\pact-ruby-e2e-example-master\provider>bundle exec rake pact:verify:foobar SPEC_OPTS='' C:/Ruby25-x64/bin/ruby.exe -S pact verify --pact-helper ./provider/spec/pact_helper.rb --pact-uri ./consumer/spec/pacts/foo-bar.json

just spits out this message.

also i tried with Pact_broker

C:\Automation\PACT\pact-ruby-e2e-example-master>bundle exec rake pact:verify:foobar_using_local_broker Publishing Foo/Bar pact to pact broker at http://localhost:9292 Error making request - Errno::ECONNREFUSED Failed to open TCP connection to localhost:9292 (No connection could be made because the target machine actively refused it. - connect(2) for "localhost" port 9292) C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact_broker-client-1.20.0/lib/pact_broker/client/pacts.rb:35:in `get', attempt 1 of 3

I get this

Well first we should target with having pact files locally

I agree @bheemreddy181 Pact is getting generated without any issue. Its just that pact:verify won't work

You have pact broker running locally? Well ignore that let?s target first with having pact files locally

sure

With out spitting out an error it?s hard to debug this

i know..

Did you just clone the above repo and running on local

yes. I just cloned, did bundle install and running these command..

commands*

My gut feel is that it's a path/escaping issue on windows.

You know the command that you see printed out?

Can you try running that directly?

this one? SPEC_OPTS='' C:/Ruby25-x64/bin/ruby.exe -S pact verify --pact-helper ./provider/spec/pact_helper.rb --pact-uri ./consumer/spec/pacts/foo-bar.json

Yes

hmm.. i got something different now

C:\Automation\PACT\pact-ruby-e2e-example-master\consumer\spec\pacts>C:/Ruby25-x64/bin/ruby.exe -S pact verify --pact-helper ./provider/spec/pact_helper.rb --pact-uri ./co nsumer/spec/pacts/foo-bar.json Traceback (most recent call last): 11: from C:/Ruby25-x64/bin/pact:23:in `<main>' 10: from C:/Ruby25-x64/bin/pact:23:in `load' 9: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/bin/pact:4:in `<top (required)>' 8: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/thor-0.20.3/lib/thor/base.rb:466:in `start' 7: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/thor-0.20.3/lib/thor.rb:387:in `dispatch' 6: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/thor-0.20.3/lib/thor/invocation.rb:126:in `invoke_command' 5: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/thor-0.20.3/lib/thor/command.rb:27:in `run' 4: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/lib/pact/cli.rb:26:in `verify' 3: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/lib/pact/cli/run_pact_verification.rb:14:in `call' 2: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/lib/pact/cli/run_pact_verification.rb:20:in `call' 1: from C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/lib/pact/cli/run_pact_verification.rb:43:in `load_pact_helper' C:/Ruby25-x64/lib/ruby/gems/2.5.0/gems/pact-1.42.3/lib/pact/cli/run_pact_verification.rb:43:in `load': cannot load such file -- ./provider/spec/pact_helper.rb (LoadError)

That's more helpful.

So, this is what I was asking you to do:

^^ This is a build I created using a free service called appveyor

This is the config file https://github.com/pact-foundation/pact-ruby-e2e-example/blob/master/.appveyor.yml

Can you run that command that you just ran again from the root of the project this time @shilpa3006

It can't find the `./provider/spec/pact_helper.rb` file because you're running it from the `spec/pacts` directory, and that's a relative path.

ok, this is annoying. When I run the command directly, it passes: https://ci.appveyor.com/project/pact-foundation/pact-ruby-e2e-example/builds/29754338

Can you try the same @shilpa3006?

`bundle exec C:/Ruby25-x64/bin/ruby.exe -S pact verify --pact-helper ./provider/spec/pact_helper.rb --pact-uri ./consumer/spec/pacts/foo-bar.json` from the directory: `C:\Automation\PACT\pact-ruby-e2e-example-master`

ok

omg

its passed :slightly_smiling_face:

C:\Automation\PACT\pact-ruby-e2e-example-master>bundle exec C:/Ruby25-x64/bin/ruby.exe -S pact verify --pact-helper ./provider/spec/pact_helper.rb --pact-uri ./consumer/s pec/pacts/foo-bar.json INFO: Reading pact at ./consumer/spec/pacts/foo-bar.json Verifying a pact between Foo and Bar A retrieve thing request with GET /thing returns a response which has status code 200 has a matching body includes headers "Content-Type" which equals "application/json" 1 interaction, 0 failures

This is an achievement for me

Thanks a lot @bethskurrie

It would be helpful if bundle exec rake:verify would work

I was hoping that command would error so I could see what was wrong :joy:

But good that it works.

I'll have a think about what could be the issue

:+1:

Now i have issue with pact-broker. somehow not able to publish to localhost:9292

Perfecto

Well pact:verify itself works fine this is an issue from windows from what I understand and you should always be on the root while running these commands

@shilpa3006 you'll need to be more specific if you want help. Please say what command you've used, and what the outcome was.

Sure @bethskurrie Before going to pact-broker, I have some questions. I have observed that BarApp's headers and payload have been hardcoded in this example under provider folder - bar_app.rb. If I want to invoke real time service, how can i do that?

@bheemreddy181: Please let me know how you invoking real app were headers and payload is dynamic?

Also, all our headers contain bearer token which generates and lives for few min

@bethskurrie @matt.fellows @bheemreddy181: I have very basic question. If provider services are in .Net, can i write my provider tests in Ruby or its mandatory that i should be writing in .Net since services are developed in .Net?

The only code you need to write for the provider side is the provider state set up code.

To set up the right state in .Net the provider verifier needs to make a request to a url. It's much easier if this url is an endpoint that gets run as part of the underlying provider application code, because you can reuse the services and db connections to insert the right data.

So, no, it's not mandatory to use Net code, but it will be much easier to manage the provider states.

Please read https://docs.pact.io/best_practices/provider/handling_auth

In fact, please read all the docs.

They're not that long.

Sure Beth. I will go through this. Thanks.

So, it means that from Pact-Ruby framework, I can invoke actual app for getting actual response. But, how can i replace this code, incase i want to invoke actual app?

```Pact.service_provider "Bar" do app { BarApp.new }```

You can't in that library.

That's for native Ruby applications. If you want to verify an app in another language, you'll need to use pact-provider-verifier.

No, sorry, you should use the native implementation that wraps the pact-provider-verifier.

Don't use the pact-provider-verifier directly.

There's one in .net

ok. so, i can use consumer tests in ruby and use pact-provider-verifier for .net? where do my provider tests be in? Sorry for basic questions. I am trying to understand.

So basically, find the Pact library that matches the language of the code base you are testing

in your case, if your consumer is Ruby use `pact-ruby` (https://github.com/pact-foundation/pact-ruby). If your back end is .NET use `pact-net` (https://github.com/pact-foundation/pact-net)

If you can?t use the language bindings for some reason, there are other approaches. For the provider side, see https://docs.pact.io/implementation_guides/other_languages#how-it-works (which includes the verifier mentioned). But only use this as a last resort

> where do my provider tests be in As a general rule, they should live with the actual provider code. It?s not necessary, but in the vast majority of cases this is where it makes sense.

Typically provider and consumer can talk two different languages but they should talk the same thing from API perspective and we can achieve that using pact

Got it. In our case consumer and provider both are in .Net. So, looks like it makes sense to implement pact-net then.

Is there a channel in slack for pact-net in case I have questions?

Of course

#pact-net

I left that as an exercise to the reader, Bheem :laughing:

Thanks Matt and Bheem.

Is the new pending pacts feature supported in pact-ruby yet?

What is this feature?

I probably cannot answer your question but I?m curious.

@joseph.haig - this one: http://blog.pact.io/p/d7c9601f-e2ca-4208-b1c0-f02eaf96c3c2/

Thanks.

@bethskurrie?

Hey team , we use our own library for authentication between backend services and it?s typically based on the current date and converts the auth to hash with using pact how can I override the authentication - do I ignore the auth and add a test in some other environment ?

Is the hash passed as a header in each request?

Is the problem that it is time based? You might be able to use the timecop gem to fix the time and so fix the authentication hash. (If you are using Rails you can use the built in time helpers instead of timecop)

and then, in the `given` condition, include a note about the time so that the service provider can do the same.

Yes hash passed as a header in each request

I will revisit this and get back with more information

I would look into not including the authorization header in the pact and instead using the ```--custom-provider-header``` option of pact-provider-verifier https://github.com/pact-foundation/pact-provider-verifier

Is the problem on the consumer side or provider (guessing provider, but want to be sure we?re on the same page)

See also https://docs.pact.io/faq#how-do-i-test-oauth-or-other-security-headers

typically I see this as an issue on both sides, on my mock i can't typically say how my Hash will look like while setting up mocked expectations for provider for example the header looks something like this ```"Authorization": "APIAuth sample:/m/EVGq2l60r13qd7azBf9qqO20="```

@matt.fellows ^ here is the sample

Thanks @simon.nizov I will give a look with custom headers basically as it?s been a valid request we will go ahead add a custom auth header you mean correct ?

The idea is not to include authentication in the pact. Only add the headers for the provider verification step.

I was not able to achieve this in the provider states ended up having it on the middleware not sure if this is a great idea , rather than mocking to return true we are actually signing the request

I would still recommend having the header in the pact (for documentation and for completeness, if not too difficult). And then on the provider verification overriding where necessary

@simon.nizov yes, but only if you're using a pactflow Pact Broker.

I don't want to do an official release of the API until we've integrated it with Pact JVM, because I don't want to have to make any breaking changes if I discover that the API that works for Pact Ruby doesn't work for Pact JVM.

If you're using your own broker you can enable the API by setting `PACT_BROKER_FEATURES=pacts_for_verification` . You'll need to grab the latest versions of the pact gems.

Thanks!

I'm just doing a release now that you'll need.

pact ruby

I'm putting out 2.48.0 of the broker and 1.47.0 of pact ruby. You'll need those two together for the informational messages to work correctly.

There's a whole lot of explanatory text that should help understand what's going on

Just be aware that at the moment, the relation that pact-ruby uses to look up the api is `beta:pacts-for-verification`.` When I make the public release, I'll be removing the `beta:` relation and changing it to `pb:pacts-for-verification` so if you had some code running in beta, and it stops working, upgrade both the broker and the gem, so that the officially released gem and API can work correctly together.

Ok thanks for the info and the heads up!

you mean a regex header during mocks ?

I just mean that you should still have the consumer code generate the correct auth header, even if it?s not correct by the time the provider verifies it (and use the approaches described to update it on demand during test cycle)

You could use regex or other matchers on the consumer side if needed

Quick question regards to pact provider verification job - from consumer side we want to verify a branch all the time, but where as from provider side we want to look at the master tagged pact. so for this setup do we setup provider as below ``` consumer_version_tags [{name: ENV['pactConsumertags'], all: false, fallback: 'master'}]``` or just use different ENV variable for each job setup ?

> consumer side we want to verify a branch all the time I don't understand what you mean by this

Typically as we are running on the PR on GitHub its the branch name or GIT SHA what we are tagging the pact with and we want to verify that particular tag all the time

Is this on a job triggered by a webhook?

or is this during the providers "normal" verification, as part of it's own build?

Yes

Webhook initiated from consumer side

You want to verify the pact URL that comes from the webhook.

You don't want to verify by the tag.

The reason is, there is metadata in that URL that will make sure your results go to the right commits.

If you make a fresh request to the broker to fetch by tag, you'll lose that.

For regular provider job we always want to look at master pact

you need two separate configurations.

one for changes made by your provider, one for when your pact changes.

So send this ${pactbroker.pactUrl} from webhook ? And there is no need for tagging

no

I would tag anyway

it will help you identify which pact belongs to which consumer branch.

Makes sense so the packURL to be shared on the master build as well ?

As it is all one webhook

Yes

How about the provider side verification job it will look at the tags ?

Or default to master all the time

I've just updated https://github.com/pact-foundation/pact_broker/wiki/Webhooks#provider-ci

see if that makes sense

It depends if you're using matching branch names between the consumer and provider

Thanks. Appreciate it we are not using matching branch names as of now

just verify what ever you call your master/dev and prod

And these are nothing but your tags tagged from consumer side on their master build if I understand this correctly

for master, yes.

for prod, you tag that at deploy time.

Please read ^^ if you haven't already

It's been updated recently, so even if you have read it, it's worth another read

Thanks Beth, Happy Tuesday :)

Hey if I am getting this correct there will two rake tasks on Ruby implementation provider in which one will be consuming the pact url from the webhook and running verification and other take task is always pulling the master tagged pact

And can we use pact-provider-verifier CLI client for running these two tasks independent of which implementation the provider is on

@bethskurrie when ever you are free ^ for couple of questions

> Hey if I am getting this correct there will two rake tasks on Ruby implementation provider in which one will be consuming the pact url from the webhook and running verification and other take task is always pulling the master tagged pact Yes.

That's what I was trying to say :stuck_out_tongue:

> And can we use pact-provider-verifier CLI client for running these two tasks independent of which implementation the provider is on Yes.

has joined #pact-ruby

has joined #pact-ruby

hi there would like to get some insight (right/wrong) on the following initiative that i am going to introduce: ? BDD/Cucumber test is often used in microservice isolation testing with by mocking the dependent services ? I am going to replace the current mock server (wiremock) with pact-jvm-server so that the "assumption" made on the mock request/response can be verified on the provider test ? Any potential downside by doing that ? Is this design violating contract testing ?

Maybe not a bad idea from what I see but if your cukes are testing one-to-one communication between consumer and provider that should be, but if it is taking a long time running through the entire application then these tests might take longer to run and don't forget there will be a provider build if the contract changes. So better to have tests around the communication rather than everything

Is there a way we can run rack middleware for each provider state ?

thanks for the feedback. yeah that's quite true as in BDD we are likely to have some test scenarios that are be redundant in term of the contract testing. Might want to suggest: ? To revise the test cases to reuse the same mock data as much as possible ? Since swaggers are used, will recommend to run the provider test with swagger validation ( to reduce provider testing time )

@bethskurrie ^ whenever you get some time

I don't know what you mean.

Different rack middleware? No.

The app is static once it's set up.

Team, can I use APIs for creating data in provider states, typically in the setup block for provider state?

Of course. The provider state function can do whatever you need it to. Most of the times it's a local in-proc setup, but there's nothing stopping your testa to reach out and hit an API

Well is it recommended to do so or better skip the total time and just insert data into dB for doing a GET

It?s hard to provide a recommendation without understanding all of the intricacies, but generally speaking, updating the database is going to be the simplest and fastest and easiest to maintain, so I?d start that way

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

Hey @bethskurrie - did you get a chance to have a look at this? :point_up:

I haven't as my OSS day isn't happening this fortnight. But I haven't forgotten it. Thanks.

Quick question folks , if i am trying to retrieve pacts with a particular tag say master like mentioned here https://github.com/pact-foundation/pact_broker/wiki/Using-tags#retrieving-pacts how does my rake task look like ? ```Pact::VerificationTask.new(:pull_request_verification) do |pact| pact.uri "#{REMOTE_PACT_BROKER_BASE_URL}/pacts/provider/provider_namee/consumer/consumer_name/latest", pact_helper: 'spec/service_consumers/pact_helper.rb' end```

how does this rake task use config from my pact_helper which is like below ```honours_pacts_from_pact_broker do # This example points to a local file, however, on a real project with a continuous # integration box, you would use a [Pact Broker](https://github.com/bethesque/pact_broker) or publish your pacts as artifacts, # and point the pact_uri to the pact published by the last successful build. pact_broker_base_url "http://pact_broker.dev.com" consumer_version_tags [{name: 'master'}] end end```

I mean how does it look for the master tag which is specified in the helper

Can some one help me with this question ?

@bethskurrie Sorry if you don?t mind can you help me getting some answer for this question ?

Hey.

Your custom task ignores the tags

You give it a specific URL.

If you have configured a custom ```app { MyApp.new }``` it will use that

it will also pick up the provider version number and tags.

But it will ignore any configurations that tell it where to get the pacts from, because the custom task overrides that.

If in that case if I don?t have my custom task how will it know which consumer it has to pick from broker ? When I say bundle exec pact:verify ?

Can I pass variables to pact verification task similar to pact publish task like the uri tags and version something like that @bethskurrie ?

It asks the broker for all the consumers for that provider.

You just provide the provider name, and the broker knows which consumers it has.

Yes, you can pass env vars to the verification task the same way you can to any ruby code.

I was more asking something like below ```PactBroker::Client::Verification.new(:Provider_Pull_Request_Verification) do | task | task.provider_version = ENV['TRAVIS_BRANCH'] task.pact_broker_base_url = REMOTE_PACT_BROKER_BASE_URL task.tags = 'master' end``` So that we can avoid something like this ```Pact::VerificationTask.new(:pull_request_verification) do |pact| pact.uri "#{REMOTE_PACT_BROKER_BASE_URL}/pacts/provider/sample/consumer/sample/latest", pact_helper: 'spec/service_consumers/pact_helper.rb' end``` and the idea here is we can run multiple provider branch level verifications in different jobs on CI rather than just running pact:verify which pulls all the pacts at once. thoughts ?

Pass in the URL of the changed pact

You still need the spec helper

What you are trying isn't supported.

All the custom verification task does is override which pact gets verified

Otherwise you'd have to duplicate your tags, versions and app configurations, potentially leaving to inconsistencies.

while consumer initiates the verification we are passing in the changed pact , this is for provider side verification for each branch change

If I have defined an interaction can I do the equivalent of `expect(WebMock).to have_requested(:post, url)` for that interaction? Or do I just depend on the fact that the test will report an error if the request has not been made?

has joined #pact-ruby

Hi guys, I have been trying to write consumer tests for a file upload api and it has to do with image/binary files. I couldn't get it to work so far and it looks like the mock service couldn't handle the binary file. I saw that there is this ticket https://github.com/pact-foundation/pact-mock_service/issues/32. Is that issue still applicable? If so, do you guys have any recommendations as far as testing image upload goes?

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

has joined #pact-ruby

according to feature_support matrix, Asychronous message pacts is not yet implemented for Ruby. Are there any timelines decided for it?

I've been hoping to get it done each week for the last 3 weeks.

I'll see what I can do next week.

has joined #pact-ruby

has joined #pact-ruby

Hey all! I'm about to implement pact as a provider in a ruby repo. I'm following the documentation here: https://github.com/pact-foundation/pact-ruby

I've added pact_helper.rb, but I'm missing what should go in line 4 here - when I run the rake task, what kicks off an actual test run?

You need to add the task as a dependency of the default rake task

Oh, no, that task is not right

I'll do an example when the kids are in bed.

Just follow that first example

Thanks @bethskurrie, that seems to be working :slightly_smiling_face:

has joined #pact-ruby

~Is this the correct channel for stub server questions?~

~The stub server seems to want an exact body match on requests made to it but we want to be flexible on one field in particular in a JSON POST body and it doesn't seem to support falling back to the matchers specified in the pact~

Hey @bethskurrie I'm following up on the above, have run into an error:

I found this info: https://www.rubydoc.info/gems/pact/1.0.19

I'm using a local file to test it's working so I tried adding the method `app_in_config_ru` to pact_helper.rb

But no change

In the "verifying pacts" docs, look at the first line in the example `Pact.service_provider` block.

Thanks @beth (http://Pactflow.io), I think that's moved us forward, we're now puzzling over another error

Trying to follow the code, it seems to be internal to the Pact gem, but I think it must be trying to run our coded block with the --pact-helper flag?

has joined #pact-ruby

@ian.mcwilliams I can't see anything unusual in that code that would explain the error you're getting.

Here's an example that I know works. https://github.com/pact-foundation/pact-ruby/blob/master/example/animal-service/spec/service_consumers/pact_helper.rb

What command are you using to run the tests?

And what is the output.

Can you checkout the pact-ruby codebase, go into `example/animal-service` run `bundle install` and then `bundle exec rake pact:verify`

Also, when you're running your failing test, can you set the env var `BACKTRACE=true` - you might get more of a stack trace.

Thanks @bethskurrie, I've downloaded pact-ruby and that's running fine

I'm using `rake test_contract` to run the tests

Here's the output from the failing test with BACKTRACE=true

has joined #pact-ruby

hi All, is there an easy way to prevent the `reports/pacts/help.md` file from being created in my repo when I verify the pacts on the provider side? I realize I can just .gitignore this but I?d rather not have to fiddle with that file

@bethskurrie just an update - I got some support in the team to look into this and as I think you expected it's a problem with spinning up the system. We tried a few things and got the setup into good shape.

It still doesn't work, as the system runs on a thin server and throws :async, which the pact gem doesn't know what to do with as it's running WEBrick

The system is using Event Machine so the server can't be replaced